www.nature.com/scientificreports OPEN Structural characterization and duplication modes of pseudogenes in plants Flavia Mascagni1, Gabriele Usai1, Andrea Cavallini1 & Andrea Porceddu2* We identifed and characterized the pseudogene complements of fve plant species: four dicots (Arabidopsis thaliana, Vitis vinifera, Populus trichocarpa and Phaseolus vulgaris) and one monocot (Oryza sativa). Retroposition was considered of modest importance for pseudogene formation in all investigated species except V. vinifera, which showed an unusually high number of retro-pseudogenes in non coding genic regions. By using a pipeline for the classifcation of sequence duplicates in plant genomes, we compared the relative importance of whole genome, tandem, proximal, transposed and dispersed duplication modes in the pseudo and functional gene complements. Pseudogenes showed higher tendencies than functional genes to genomic dispersion. Dispersed pseudogenes were prevalently fragmented and showed high sequence divergence at fanking regions. On the contrary, those deriving from whole genome duplication were proportionally less than expected based on observations on functional loci and showed higher levels of fanking sequence conservation than dispersed pseudogenes. Pseudogenes deriving from tandem and proximal duplications were in excess compared to functional loci, probably refecting the high evolutionary rate associated with these duplication modes in plant genomes. These data are compatible with high rates of sequence turnover at neutral sites and double strand break repairs mediated duplication mechanisms. Pseudogenes are genic derived sequences that have lost the capability to encode a functional protein1–3. Based on the intron-exon structure they are classifed into two major types: non processed (or duplicated) and processed (or retroposed). Non processed pseudogenes arise from genome or chromosomal duplications and usually retain the exon-intron structure of ancestral genes 4. Processed pseudogenes derive by genomic integration of reverse transcribed messenger RNA, are typically devoid of introns, presenting fanking direct repeats and a poly-adenine tail at their 3’ end2,5. Te functional paralogous showing the top sequence and structural homology to a pseu- dogene is referred to as “parental” locus or the youngest pseudogene prototype3,6. Although defunct copies of functional genes, pseudogenes have recently stimulated a wide range of scientifc interests ranging from structural and evolutionary genomics7 to post-transcriptional gene expression regulation studies 3,8,9. Comparative analyses of the pseudogenes have proved informative for both genome and gene evolutionary studies. For example, the balance between pseudogenes types has indicated diferent modes of gene amplifcation between plants and mam- malians. Te high prevalence of processed over non processed pseudogenes in mammalians genomes has been proved in sharp contrast with the low ratios of processed versus non processed types in plants 10,11. Non collinear sequence duplication would occur mainly by retro-positions in mammalians3,11 while double strand break repair connected mechanisms would be prevalent in plants10,12. Pseudogenes have been used also to infer past genomic organization and gene family evolution. Te high abundance of pseudogenes belonging to specifc gene families has been used as an indication of low selective pressure on genic functions that are no longer important for the ftness of one species. To cite an example, H. sapiens presents a high number of olfactory receptor pseudogenes indicating that, during evolution, the functions exerted by these genes have lost importance in the hominids 13. However, in other cases, pseudogene copy number correlates with the expression level of the functional parental gene14. Intraspecifc analysis of pseudogenes have revealed how they contribute to genome plasticity by favor- ing sequence diversity. Indeed, as a consequence of low selective pressure, pseudogenes show a high propensity to accumulate mutations which can be transferred to functional genes by gene conversion mediated events 1. Zhu et al. 15 have demonstrated that conversions involving pseudogenes could accelerate gene family evolution and divergence relative to their orthologs. Gene conversion has been cited by Prade et al. 10 to explain the posi- tive association between physical distance and sequence diversity of pseudogene parental-pairs. Te interest 1Department of Agricultural, Food, and Environment, University of Pisa, Via del Borghetto 80, 56124 Pisa, Italy. 2Dipartimento di Agraria, Università degli studi di Sassari, Via Enrico de Nicola 1, 07100 Sassari, Italy. *email: [email protected] Scientifc Reports | (2021) 11:5292 | https://doi.org/10.1038/s41598-021-84778-6 1 Vol.:(0123456789) www.nature.com/scientificreports/ Figure 1. Procedures followed to identify and classify the pseudogenes based on their inferred intron-exon structures. DUP duplication, RET retroposition, AMB ambiguous. for pseudogene has recently gained new emphasis following the demonstrations that they may play a role in regulating the expression of their functional cognate16,17. A recent study on seven plant genomes showed that a surprisingly large fraction of non transposable element (TE) and regulatory non coding RNAs (microRNA and long non coding RNAs) originate from transcription of pseudogene proximal upstream regions 7. Based on these observations it was proposed that the recruitment of pseudogene-derived regulatory sequences may lead to the origin of novel regulatory modules7. Comparative analyses of pseudogene complements have been demonstrated very informative in illustrating how remodelling processes mark each organism genome evolution. For example, the comparison of pseudogene complements across H. sapiens, C. elegans and D. melanogaster phyla highlighted lineage specifc features capable of refecting genome histories more than the conservation of essential biological functions18. Specifcally, the human pseudogene complements illustrated a massive burst of retroposition activ- ity at the dawn of the primates while the fy and worm refected a history of deactivated duplications. Most of our knowledge on plant genome evolution has been gained through comparative analysis of structural features of functional gene complements, but yet very little emphasis has been dedicated to the analysis of pseudogene complements. Exploiting the genomic resources publicly available, here we present a comparison of structural features of the pseudogene complements of fve species: four dicots (A. thaliana, P. vulgaris, P. trichocarpa, V. vinifera) and one monocot (O. sativa). Te four dicots were chosen to represent diferent vegetative habitus and lifespan of the two main groups of rosids: Vitales and eurosids. V. vinifera is representative of the order Vitales, and shows a woody habitus and a perennial lifespan. Te other species belong to the main orders of eurosids: the Fabidae with the annual and herbaceous species P. vulgaris, the Malvidae with the perennial and woody P. trichocarpa and the annual and herbaceous A. thaliana. Our analysis confrmed that processed are less abundant than non processed pseudogenes in all investigated species but highlighted also a signifcant variation for the relative abundance and genomic distribution of pseudogene types which refected lineage specifc diferences. Furthermore, although the majority of plant pseudogenes seems to derive from duplicative transpositions, the analysis of pseudogene-parental locus genomic positions refected evolutionary diferences for the prevalent modes of pseudogene origin among the investigated species. Results Pseudogenes were identifed based on sequence homology to functional loci in Arabidopsis thaliana, Populus trichocarpa, Phaseolus vulgaris, Vitis vinifera and Oryza sativa genomes. Te pipeline used for pseudogene identifcation and classifcation is represented in Fig. 1. Translated coding exons of each functional locus were Scientifc Reports | (2021) 11:5292 | https://doi.org/10.1038/s41598-021-84778-6 2 Vol:.(1234567890) www.nature.com/scientificreports/ Figure 2. Numbers and distributions of pseudogenes in the fve investigated species. (a) Total numbers of pseudogenes identifed in each species. (b) Pseudogene distributions based on their positions in either DNA strands. UTR untrnslated 5’, 3’ exons of genes and introns therein, Intron_cds intron of the coding sequences, Intergenic pseudogenes that map between two consecutive genes. used to query the hard-masked genomic sequences by tBlastN 19. Hits not overlapping coding sequence coor- dinates and matching consecutive exons of query loci were merged to identify putative pseudogenes. When a genomic region matched overlapping pseudogene models, only the model with the highest homology was chosen for further analysis (Fig. 1). Tis criterion derives from the hypothesis that the functional homologous showing the highest sequence identity to the pseudogene sequence is the best extant version of the parental locus6. Pseudogene models were predicted by aligning the query protein (encoded by the inferred parental locus) to the pseudogene genomic sequence. When introns were predicted at all the positions expected based on the model of the query locus, the pseudogenes were classifed as non processed. If no introns were predicted, though expected based on the parental gene model, the pseudogenes were classifed as processed. Pseudogenes with models showing features of both duplicated and processed